GCC Bugzilla Bug 38593

seen with 20081213, on hppa-linux-gnu, building with -g -O2 -fPIC

/home/packages/openjdk/openjdk-6-6b14~pre1/openjdk-ecj/hotspot/src/share/vm/interpreter/bytecodeInterpreter.cpp:
In static member f
unction 'static void BytecodeInterpreter::run(BytecodeInterpreter*)':
/home/packages/openjdk/openjdk-6-6b14~pre1/openjdk-ecj/hotspot/src/share/vm/interpreter/bytecodeInterpreter.cpp:402:
error: goto de
stination is neither a label nor a pointer
goto D.98183;

/home/packages/openjdk/openjdk-6-6b14~pre1/openjdk-ecj/hotspot/src/share/vm/interpreter/bytecodeInterpreter.cpp:402:
error: goto de
stination is neither a label nor a pointer
goto D.98183;

[this message repeats]

/home/packages/openjdk/openjdk-6-6b14~pre1/openjdk-ecj/hotspot/src/share/vm/interpreter/bytecodeInterpreter.cpp:402:
internal compiler error: verify_gimple failed
Please submit a full bug report,
with preprocessed source if appropriate.

Created an attachment (id=16952) [edit]
preprocessed source

Created an attachment (id=16953) [edit]
build log

precompiled header file at:
http://people.ubuntu.com/~doko/tmp/_precompiled.incl.gch.bz2

Does disabling PCH fix the issue (just adding -D__disable_PCH__ should be
enough to disable PCH when building that one file)?

no, same result.

btw, in the generated .ii file using -D__disable_PCH__ -save-temps, the
precompiled header is still referenced.

That's unexpected.  Anyway, if you move away the gch file (or directory), can
you still reproduce it and if yes, attach preprocessed source?

Otherwise it would be extremely hard to reproduce for anyone...

Maybe the bug boils down to PR38725.

Created an attachment (id=17034) [edit]
preprocessed source

That's really a duplicate of PR38725: The code contains many lines of the form

  goto *dispatch_table[opcode];

where dispatch_table[opcode] is of the type uintptr_t which is a typedef
to unsigned int. Writing

  goto *(void*)dispatch_table[opcode];

instead makes the ICE disappear.


*** This bug has been marked as a duplicate of 38725 ***